This is great. I write 100-200 page functional spec documents at a vendor for large scale file based broadcast systems (10-300 linux servers) for a number of c

This is great. I write 100-200 page functional spec documents at a vendor for large scale file based broadcast systems (10-300 linux servers) for a number of customer projects, and have been trying to get away from word since it's slow at that scale and I don't want to spend any time on formatting, and want to get to a templated approach for the others on my team with a consistent output. Currently I'm using markdown with some CSS and just use Marked2 (Mac) to export but don't have it all worked out yet. Markdown + LaTeX + Pandoc is probably better and more powerful or precise than using CSS. I don't use equations but I do use tables a lot and I'm using multimarkdown ascii ones for now (with a nice atom.io auto-format plugin to make it easy to author) and some code blocks with syntax highlighting. The idea is to have a folder per customer/project spec, with a consistent structure of one .md file for the body and a local sub-folder for images (svg workflow(bpmn)/system diagrams mostly, some jpgs for logos, screenshots). The folder would be in a local git repo so we can commit changes and export diffs to see what changed between versions and have multiple people work on the same doc with tracking. I'm using "invisible links" for in-line comments at the bottom of each section that are added while going though it with the customer since it usually takes 5-30 versions before it's finalized and signed off. Those keep a record of discussion with the customer and don't get rendered out in the final output. Also using standard set of status tags (@outstanding, @done, @info) within the comment text. Ex; `[Note: <initials> YYY-MM-DD]: # (comment text @status-tag)`. Going through the in-progress spec with customers and typing notes inline has been much better than word's commenting system and using markdown makes it easy for the customer to read without extraneous formatting code in-line. Currently using a multimarkdown header for variables; customer name, project name, author, author email, spec version, etc., but I might move that to a separate YAML file. Ideally, to make each version of the spec it would be markdown through LaTeX/Pandoc to render a PDF with; - Title page generated automatically using variables (multi markdown header or separate YAML) - Automatic Table of Contents - Automatic header numbering (h1-h6) - Automatic header/footer using variables & auto page numbering - Ability for basic control of image size; 80% width (svg), original pixel resolution (pngs), etc., positioning. - Ability to have global paragraph numbering in sidebar that the customer can reference while discussions are ongoing, and turn that off for final output to PDF. I'm going to spend time with the examples from the original link to try to work that all out but any suggestions or tips would be greatly appreciated. I'd be happy to post an example of the final template and write-up of the approach on GitHub.